Dielectric Nanoprism-Based Plasmonic Sensor with Normal‑Incidence SPR for Enhanced Sensitivity

Background

Surface plasmon resonance (SPR) sensors have become a cornerstone of label‑free optical detection in biomedical and material‑science applications. Traditional designs rely on angular or spectral interrogation, with the Kretschmann configuration being the most widely adopted approach [1–5]. In this scheme, light is coupled into a thin metal film via a high‑index prism, requiring a precise incident angle to satisfy the wave‑vector matching condition:

\[\frac{2\pi}{\lambda}\,n_P\sin\theta_r = \text{Re}\big[\beta^{\text{SP}}\big]\]

where \(n_P\) is the prism’s refractive index and \(\beta^{\text{SP}}\) is the surface‑plasmon propagation constant. The need for a large incidence angle often limits the sensor’s practicality and integration potential.

Several normal‑incidence strategies have been proposed, including narrow metallic grooves [14], nanocavities [15], and dielectric gratings [16]. These designs exploit spectral dips with sub‑nanometer widths, improving the figure of merit (FOM). However, fabrication constraints—such as 3 nm groove widths—can impede scalability. An alternative is to use dielectric structures that funnel light toward the metal–dielectric interface. High‑aspect‑ratio dielectric gratings (HARDG) guide light into active layers of photovoltaics [19] and can be repurposed for sensing applications.

In this work, we introduce a nanoprism array that directs normally incident light toward a metal–dielectric interface, enhancing the evanescent field at the sensing surface. The design is fully compatible with standard lithography and spin‑coating techniques, making it attractive for rapid prototyping and large‑scale production.

Methods

The device geometry is illustrated in Fig. 1a. A periodic array of isosceles nanoprisms is etched into a MgF₂ substrate. The triangular grooves are filled with aluminum zinc oxide (AZO), which can be deposited via spin‑coating, yielding a planar surface for subsequent metal deposition. A thin gold film (30 nm) is then sputtered onto the AZO, forming the plasmonic layer. Water is used as the analyte to emulate biosensing conditions. Optical constants were taken from reputable databases: MgF₂ [22], AZO [23], and gold [24].

a Schematic of the nanoprism array; b Time‑averaged power flow at \(\lambda=758\,\text{nm}\) for the structure without the metal layer, illustrating the funneling effect.

By illuminating the structure with a TM‑polarized plane wave from the substrate side (field amplitude 1 V/m), the simulated power flow (Fig. 1b) confirms that the nanoprisms focus energy onto the metal–dielectric interface, enhancing the local field relative to the classic Kretschmann configuration. The resonance wavelength is tunable via the prism period, P.

The device is modeled in COMSOL Multiphysics using the finite‑element method. Validation against analytical solutions for the Kretschmann setup confirms the model’s accuracy. Optimization targets two key metrics: (i) field amplitude at the metal–water interface and (ii) full‑width‑at‑half‑maximum (FWHM) of the reflectance dip. Sensitivity (S_B) and FOM are defined as:

S_B = \frac{\Delta\lambda}{\Delta n} \quad\text{(nm/RIU)}\quad\text{(Eq. 2)}\nFOM = \frac{S_B}{\text{FWHM}} \quad\text{(1/RIU)}\quad\text{(Eq. 3)}

Optimization yields the following geometrical parameters: buffer thickness t_BL = 100 nm, metal thickness t_M = 30 nm, prism width w_G = 325 nm, height H = 700 nm, and period P = 550 nm. These values balance performance with realistic fabrication tolerances.

a Electric‑field magnitude map at \(\lambda=758\,\text{nm}\) for TM polarization; b Field decay along the propagation direction for Kretschmann (black dashed) and nanoprism (red solid) configurations.

Spectral reflectance curves (Fig. 3a) show that the nanoprism device maintains a sharp resonance even as the analyte refractive index approaches that of the buffer layer. The maximum sensitivity achieved is 250 nm/RIU with an FOM of 100 1/RIU—values that surpass typical Kretschmann results [26–30].

Results and Discussions

To broaden the operational refractive‑index range, the buffer layer was replaced from AZO to gallium phosphide (GaP) (optical constants from [34]). GaP’s higher index mitigates the spectral‑peak degradation observed with AZO when the analyte index increases. The nanoprism material remains AZO to preserve the funneling effect.

Silver offers superior plasmonic performance compared to gold, but its lack of biocompatibility limits its use in biosensing. We therefore explored a bimetallic stack: 25 nm Ag coated with 5 nm Au. Fig. 4a demonstrates that this bilayer produces a resonance that combines the sharpness of silver with the protective layer of gold. The optimized stack yields a spectral dip at a shorter wavelength and a narrower FWHM compared to either metal alone.

a Reflectance spectra for single‑metal (gold, silver) and bimetallic (Ag/Au) configurations; the arrow indicates the optimum 25 nm Ag/5 nm Au stack.
b Reflectance for the GaP‑buffered, bimetallic device across various analyte refractive indices—peak sharpness is preserved.
c Sensitivity and FOM versus refractive index for the optimized design; the vertical line marks the limit of the AZO‑buffered, gold‑only case.

With the GaP buffer and Ag/Au bilayer, the sensor attains a maximum sensitivity of 450 nm/RIU that remains stable across a refractive‑index range of 1.33–1.43. Correspondingly, the FOM spans 160–220 1/RIU—outperforming recent graphene, silicon, oxide, and metallic nanoprisms approaches [27–37].

Conclusions

We have demonstrated a dielectric nanoprism sensor that amplifies the local field at the metal–analyte interface, enabling deeper SPR penetration and larger interaction volumes. The normal‑incidence architecture simplifies integration—e.g., with fiber‑optic probes—and delivers exceptional performance: a plateau sensitivity of ~450 nm/RIU and FOM values between 160 and 220 1/RIU over 1.33–1.43 RIU. Material optimizations—including a GaP buffer and an Ag/Au bilayer—extend the operational index range while maintaining biocompatibility. These attributes make the device suitable for biomedical diagnostics, environmental monitoring, and industrial liquid sensing.